known to control differentiation of root hair cells (21, 22), the cell type infected by Rhizobium. Thus, endogenous ethylene may affect the persistence of rhizobial infection by controlling the formation of infectable root hair cells. Alternatively, ethylene may act as a diffusible signal for activation of mechanisms that arrest rhizobial infection. ACC is inhibitory to nodulation when ap- plied after the initiation of rhizobial infec- tion (Fig. 5A, 24 and 48 hours). Similarly, the ethylene biosynthesis inhibitor AVG can increase nodule number when applied after the initiation of infection (5). These observations are consistent with a model wherein endogenous ethylene acts subse- quent to infection initiation and root hair differentiation. If ethylene provides a signal for induc- tion of infection arrest, then plant cells containing persistent Rhizobium infections either must avoid the ethylene signal or must be insensitive to the signal. Localized production of ethylene at sites of infection arrest could facilitate avoidance of ethylene by infections destined for nodule coloniza- tion. A model for cell-specific regulation of ethylene synthesis during root hair cell dif- ferentiation has been proposed in Arabidop- sis (22). In wild-type M. truncatula, all rhi- zobial infections can be blocked by treat- ment with ACC as late as 48 hours after inoculation (Fig. 5A), indicating that infec- tions destined for nodule colonization are not inherently insensitive to ethylene. However, after macroscopic nodule primor- dia appear, nodulation is largely insensitive to exogenous ACC (Fig. 5A, 72 hours); thus, sustained rhizobial infections may ac- quire insensitivity to ethylene. In plant-pathogen interactions, ethylene has been implicated as an endogenous cue for induction of host defense-related genes (23). Despite extensive correlative data, however, a causal role for ethylene in resist- ance to pathogens has not been established (24). In M. truncatula, the sickle mutation causes extensive developmental abnormali- ties and hyperinfection by Rhizobium, which indicates that skl1 encodes a function com- mon to both plant development and con- trol of rhizobial infection. REFERENCES AND NOTES ___________________________ 1. T. Bhuvaneswari, A. A. Bhagwat, W. D. Bauer, Plant Physiol. 68, 1144 (1981); G. Caetano-Anolles and W. D. Bauer, Planta 175, 546 (1988); G. Caetano-Anolles and P. M. Gresshoff, J. Plant Physiol. 138, 765 (1991). 2. B. J. Carroll, D. L. McNeil, P. M. Gresshoff, Plant Physiol. 78, 34 (1985); Proc. Natl. Acad. Sci. U.S.A. 82, 4162 (1985). 3. J. Vasse, F. de Billy, G. Truchet, Plant J. 4, 555 (1993). 4. K. Niehaus, D. Kapp, A. Puhler, Planta 190, 415 (1993); S. Perotto, N. J. Brewin, E. L. Kannenberg, Mol. Plant-Microbe Interact. 7, 99 (1994). 5. N. K. Peters and D. K. Crist-Estes, Plant Physiol. 91, 690 (1989). 6. J. C. Fearn and T. A. LaRue, ibid. 96, 239 (1991); F. C. Guinel and T. A. LaRue, ibid. 97, 1206 (1991); ibid. 99, 515 (1992). 7. X. Ruan and N. K. Peters, J. Bacteriol. 174, 3467 (1992). 8. K. H. Lee and T. A. LaRue, Plant Physiol. 100, 1759 (1992). 9. F. Ligero et al., ibid. 97, 1221 (1991); K. H. Lee and T. A. LaRue, ibid. 100, 1334 (1992). 10. R. V. Penmetsa and D. R. Cook, in preparation. 11. S. A. Leong, P. H. Williams, G. S. Ditta, Nucleic Acids Res. 13, 5965 (1985). 12. C. Boivin et al., Plant Cell 2, 1157 (1990). X-Gal is 5-bromo-4-chloro-3-indolyl--D-galactopyranoside. 13. Infections and nodule primordia are first evident on wild-type and sickle roots by 36 to 48 hours after inoculation. However, nodule development is retard- ed in sickle, such that 21-day-old nodules are about one-tenth the size of wild-type nodules of similar age. 14. M. B. Lanahan et al., Plant Cell 6, 521 (1994). 15. J. R. Ecker, Science 268, 667 (1995). 16. L. I. Knight, R. C. Rose, W. Crocker, ibid. 31, 635 (1910); P. Guzman and J. R. Ecker, Plant Cell 2, 513 (1990). 17. A. B. Bleecker, M. A. Estelle, C. Somerville, H. Kende, Science 241, 1086 (1988). 18. J. Hua, C. Chang, Q. Sun, E. M. Meyerowitz, ibid. 269, 1712 (1995); C. Chang and E. M. Meyerowitz, Proc. Natl. Acad. Sci. U.S.A. 92, 4129 (1995). 19. G. Roman et al., Genetics 139, 1393 (1995). 20. Isolated leaves were floated on water and treated with ethylene (100 ppm) (17). After 5 days dark incu- bation at 21°C, leaves of wild-type and sickle con- trols were chlorotic and green, respectively. 21. J. D. Masucci and J. W. Schiefelbein, Plant Cell 8, 1505 (1996); Plant Physiol. 106, 1335 (1994). 22. M. Tanimoto, K. Roberts, L. Dolan, Plant J. 8, 943 (1995). 23. J. R. Ecker and R. W. Davis, Proc. Natl. Acad. Sci. U.S.A. 84, 5202 (1987); F. Mauch, J. B. Meehl, A. J. Staehelin, Planta 186, 367 (1992). 24. A. F. Bent et al., Mol. Plant-Microbe Interact. 5, 372 (1992); P. Silverman et al., ibid. 6, 775 (1993); K. Lawton et al., ibid. 8, 863 (1995). 25. F. deBilly et al., Plant J. 1, 27 (1991). 26. N. Chaubet, B. Clement, C. Gigot, J. Mol. Biol. 225, 569 (1992). 27. P. Somasegaran and H. J. Hoben, Handbook for Rhizobia: Methods in Legume-Rhizobium Technolo- gy (Springer-Verlag, New York, 1994), pp. 392–398. 28. Supported by grants from the NSF Program on Inte- grative Plant Biology (IBN-9507535) and from the Samuel Roberts Noble Foundation. 15 July 1996; accepted 2 December 1996 Measuring Serotonin Distribution in Live Cells with Three-Photon Excitation S. Maiti, Jason B. Shear,* R. M. Williams, W. R. Zipfel, Watt W. Webb† Tryptophan and serotonin were imaged with infrared illumination by three-photon ex- citation (3PE) of their native ultraviolet (UV ) fluorescence. This technique, established by 3PE cross section measurements of tryptophan and the monoamines serotonin and dopamine, circumvents the limitations imposed by photodamage, scattering, and in- discriminate background encountered in other UV microscopies. Three-dimensionally resolved images are presented along with measurements of the serotonin concentration (50 mM) and content (up to 5  10 8 molecules) of individual secretory granules. Neurotransmitter granules have typically been studied either with various imaging techniques (1, 2) that do not directly detect the granular content, or with chemical or electrical assays (3, 4) that identify the granule contents but can probe only the extracellular medium. Thus, it has not been possible to determine neurotransmitter con- centration or total neurotransmitter con- tent of individual granules in intact cells. As a solution, we have excited the native UV fluorescence of these molecules by si- multaneous absorption of three infrared photons, which accesses shorter wavelength UV transitions in living cells than conven- tional or two-photon microscopy (5). When subjected to a high-intensity irra- diation at wavelength , a molecule that ordinarily absorbs at /3 may exhibit fluo- rescence at /3 by a three-photon absorp- tion mechanism (6). The average fluores- cence photon count rate F measured from three-photon excitation depends on the three-photon molecular absorption cross section  3 [unit: (length) 6 (time) 2 (pho- ton) -2 ], the instantaneous intensity I, the fluorescence quantum efficiency Q, the de- tection efficiency K, the wavelength , and the concentration C (number of molecules per unit volume). Analogous with two-pho- ton excitation (7), F is determined by the product of I 3 and C, integrated over space (r) and time (t'): F = KQ 3 [(1/t)  0 t  0  I 3 (r,t')C(r,t')dr dt'] (1) For excitation of a homogeneous dye solu- tion with a focused and pulsed laser beam with a gaussian temporal and spatial profile, the integral yields (SI units) School of Applied and Engineering Physics, Cornell Uni- versity, Ithaca, NY 14853, USA. All authors contributed equally to this work. *Present address: Department of Chemistry and Bio- chemistry, University of Texas, Austin, TX 78712, USA. †To whom correspondence should be addressed. SCIENCE  VOL. 275  24 JANUARY 1997 530 on March 25, 2019 http://science.sciencemag.org/ Downloaded from